therapy. leif oltedal, md, phd, katherine l. narr, phd ... psykisk helsevern/document… ·...
TRANSCRIPT
Accepted Manuscript
Volume of the human hippocampus and clinical response following electroconvulsivetherapy.
Leif Oltedal, MD, PhD, Katherine L. Narr, PhD, Christopher Abbott, MD, AmitAnand, MD, Miklos Argyelan, MD, Hauke Bartsch, PhD, Udo Dannlowski, MD, PhD,Annemieke Dols, MD, PhD, Philip van Eijndhoven, MD, PhD, Louise Emsell, PhD,Vera Jane Erchinger, Randall Espinoza, MD, MPH, Tim Hahn, PhD, Lars G. Hanson,PhD, MSc, Gerhard Hellemann, PhD, Martin Balslev Jorgensen, MD, DMSc, UteKessler, MD, PhD, Mardien L. Oudega, MD, PhD, Olaf B. Paulson, MD, DMSc, RonnyRedlich, PhD, Pascal Sienaert, MD, PhD, Max L. Stek, MD, PhD, Indira Tendolkar,MD, PhD, Mathieu Vandenbulcke, MD, PhD, Ketil J. Oedegaard, MD, PhD, Anders M.Dale, PhD
PII: S0006-3223(18)31534-8
DOI: 10.1016/j.biopsych.2018.05.017
Reference: BPS 13551
To appear in: Biological Psychiatry
Received Date: 7 February 2018
Revised Date: 29 April 2018
Accepted Date: 13 May 2018
Please cite this article as: Oltedal L., Narr K.L, Abbott C., Anand A., Argyelan M., Bartsch H.,Dannlowski U., Dols A., van Eijndhoven P., Emsell L., Erchinger V.J., Espinoza R., Hahn T., HansonL.G, Hellemann G., Jorgensen M.B., Kessler U., Oudega M.L, Paulson O.B, Redlich R., SienaertP., Stek M.L, Tendolkar I., Vandenbulcke M., Oedegaard K.J & Dale A.M, Volume of the humanhippocampus and clinical response following electroconvulsive therapy., Biological Psychiatry (2018),doi: 10.1016/j.biopsych.2018.05.017.
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service toour customers we are providing this early version of the manuscript. The manuscript will undergocopyediting, typesetting, and review of the resulting proof before it is published in its final form. Pleasenote that during the production process errors may be discovered which could affect the content, and alllegal disclaimers that apply to the journal pertain.
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPT
Oltedal et al. 1
Volume of the human hippocampus and clinical response following
electroconvulsive therapy.
Leif Oltedal, MD, PhDa,b,c,d, Katherine L Narr, PhDe, Christopher Abbott, MDf, Amit Anand,
MDg, Miklos Argyelan, MDh, Hauke Bartsch, PhDb,c, Udo Dannlowski, MD, PhDi, Annemieke
Dols, MD, PhDj, Philip van Eijndhoven, MD, PhDk, Louise Emsell, PhDl, Vera Jane
Erchingera, Randall Espinoza, MD, MPHe, Tim Hahn, PhDi, Lars G Hanson, PhD, MScm,n,
Gerhard Hellemann, PhDe, Martin Balslev Jorgensen, MD, DMSco, Ute Kessler, MD, PhDa,p,
Mardien L Oudega, MD, PhDj, Olaf B Paulson, MD, DMScn,q, Ronny Redlich, PhDi, Pascal
Sienaert, MD, PhDl, Max L Stek, MD, PhDj, Indira Tendolkar, MD, PhDk, Mathieu
Vandenbulcke, MD, PhDl, Ketil J Oedegaard, MD, PhDa,p,r & Anders M Dale, PhDb,c,s
a) Department of Clinical Medicine, University of Bergen, Bergen, Norway
b) Center for Multimodal Imaging and Genetics, University of California, San Diego, La Jolla, CA, USA
c) Department of Radiology, University of California, San Diego, La Jolla, CA, USA
d) Center for Medical Visualization, Department of Radiology, Haukeland University Hospital, Bergen, Norway
e) Departments of Neurology, Psychiatry and Biobehavioral Sciences, University of California, Los Angeles
(UCLA), CA, USA
f) Department of Psychiatry, University of New Mexico School of Medicine, Albuquerque, USA
g) Cleveland Clinic, Center for Behavioral Health, Cleveland, USA
h) Center for Psychiatric Neuroscience at the Feinstein Institute for Medical Research, New York, USA
i) Department of Psychiatry, University of Münster, Germany
j) Department of Old Age Psychiatry, VUmc Amsterdam/GGZinGeest, Amsterdam, Netherlands and
Neuroscience Campus, Amsterdam, the Netherlands
k) Donders Institute for Brain, Cognition and Behavior, Department of Psychiatry, Nijmegen, Netherlands
l) KU Leuven, University Psychiatric Center KU Leuven, Leuven, Belgium.
m) Center for Magnetic Resonance, DTU Elektro, Technical University of Denmark, Kgs. Lyngby, Denmark
n) Danish Research Center for Magnetic Resonance, Center for Functional and Diagnostic Imaging and
Research, Copenhagen University Hospital, Hvidovre, Denmark
o) Psychiatric Center Copenhagen, Copenhagen, Denmark
p) Division of Psychiatry, Haukeland University Hospital, Bergen, Norway
q) Neurobiology Research Unit, Rigshospitalet, University of Copenhagen, Denmark
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPT
Oltedal et al. 2
r) K.G. Jebsen Centre for Research on Neuropsychiatric Disorders, Bergen, Norway.
s) Department of Neurosciences, University of California, San Diego, La Jolla, CA, USA
Correspondence: Leif Oltedal, address: Department of Clinical Medicine, University of
Bergen, Bergen, Norway. email: [email protected] phone: +47 93044829
Key words: ECT, brain, depression, neuroimaging, antidepressant response, biomarker
Word count: Abstract; 248, article body; 3666
Number of figures: 2
Number of tables: 1
Supplemental information: 1
Short title: Hippocampal volume and response after ECT
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPT
Oltedal et al. 3
Abstract
Background: Hippocampal enlargements are commonly reported following
electroconvulsive therapy (ECT). To clarify mechanisms, we examined if ECT-
induced hippocampal volume change relates to dose (number of ECT sessions and
electrode placement) and acts as a biomarker of clinical outcome.
Methods: Longitudinal neuroimaging and clinical data from ten independent sites
participating in the Global ECT-MRI Research Collaboration were obtained for mega-
analysis. Hippocampal volumes were extracted from structural MR images, acquired
before and after patients (n=281) experiencing a major depressive episode
completed an ECT treatment series using right unilateral (RUL) and bilateral (BL)
stimulation. Untreated non-depressed controls (n=95) were scanned twice.
Results: The linear component of hippocampal volume change was 0.28%, 0.08 SE,
per ECT session, p<0.001. Volume change varied by electrode placement in the left
(BL: 3.3 ± 2.2%, d=1.5; RUL: 1.6 ± 2.1%, d=0.8; p<0.0001), but not the right
hippocampus (BL: 3.0 ± 1.7%, d=1.8; RUL: 2.7 ± 2.0%, d=1.4; p=0.36,). Volume
change for electrode placement per ECT session varied similarly by hemisphere.
Individuals with greater treatment-related volume increases had poorer outcomes
(MADRS change -1.0, 0.35 SE, per 1% volume increase, p=0.005), although effects
were not significant after controlling for ECT number (slope: -0.69, 0.38 SE,
p=0.069).
Conclusions: The number of ECT sessions and electrode placement impacts the
extent and laterality of hippocampal enlargement, but volume change is not
positively associated with clinical outcome. Results suggest the high efficacy of ECT
is not explained by hippocampal enlargement, which alone, might not serve as a
viable biomarker for treatment outcome.
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPT
Oltedal et al. 4
Introduction
Major depression is the leading cause of disability worldwide (1), yet standard
treatments for depression are only moderately successful (2). There is thus a need
to better understand the mechanisms of successful response to antidepressant
therapies, which may then inform more effective treatment interventions for patients
with major depression. Though depression is typically treated with different forms of
psycho- or pharmacotherapies, electroconvulsive therapy (ECT) is still regarded as
the most effective acute treatment for severe and treatment resistant major
depressive episodes (3). With ECT, electrical current is applied through scalp
electrodes, intentionally inducing a seizure, typically 2-3 times per week. When
administered with modern techniques under anesthesia, ECT is well tolerated and
has a good safety record. Yet, despite its safety and efficacy (3), the neurobiological
underpinnings of ECT response, as with other forms of antidepressant treatment,
remain unclear. Establishing objective biomarkers of clinical response could allow for
the timely implementation of alternative treatment strategies in unresponsive
patients.
Most neuroimaging studies of ECT demonstrate treatment-related volume increase
of the hippocampus (4-9), which suggests that hippocampal volume may serve as a
biomarker of clinical response. These observations together with data from
preclinical studies are taken as evidence to support the neurogenic theory of
depression (10). In particular, translational models provide evidence to suggest that
a decrease of adult neurogenesis in the hippocampus is associated with depression
and can potentially be reversed with ECT (10-12). This hypothesis is supported by
observations that the human hippocampus harbors neuronal stem cells that
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPT
Oltedal et al. 5
proliferate throughout life (13), that the volume of the hippocampus is frequently
reported as reduced in depression (14), and that in an animal model of ECT, a dose-
dependent increase in neurogenesis is seen (15). However, the mechanisms
underlying ECT-related volume enlargement of the human hippocampus remain
unclear, and associations with clinical outcome have not been demonstrated
conclusively (9).
In ECT practice, the number of treatments in an ECT index series typically depends
on severity of depression and the speed of recovery, such that unresponsive patients
tend to receive more ECT sessions on average (16). Bilateral (BL) electrode
placement is widely used for stimulation. However, to mitigate risk for cognitive side
effects, particularly for verbal and retrograde autobiographical memory, the use of
other electrode montages are also standard practice (17-19). In particular, right
unilateral (RUL) ECT, which was developed in an effort to reduce the spread of
seizure activity to brain areas such as left temporal cortex important for verbal
memory, is often used as a first line form of ECT (17, 18). Computational modelling
of electric fields supports that bilateral ECT leads to more diffuse brain stimulation
than more focal RUL ECT (20, 21). Both the number of ECT sessions received and
electrode placement may thus impact the extent and laterality of hippocampal
neuroplasticity and in turn the mechanisms of treatment response. However, prior
studies have lacked the sample sizes and statistical power needed to investigate the
moderating effects of these parameters, or have simply controlled for these factors
as nuisance variables. Consequently, no clear associations between dose or mode
of electrode placement and measured hippocampal structural changes have
emerged (12, 22, 23).
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPT
Oltedal et al. 6
To address the clinical relevance of ECT-related hippocampal volume change, we
included 281 patients from the Global ECT-MRI Research Collaboration (GEMRIC)
(24), and analyzed volume changes of the hippocampus after serial ECT treatment.
With the largest and most geographically diverse sample to date, and by using an
optimized image processing pipeline, we obtained sufficient statistical power to
probe for relationships between hippocampal volume, dose response (number of
sessions as well as electrode placement) and symptom improvement of relatively
small effects (24) (f2 = 0.03, α=.05, Power=.80, as estimated for a linear model with
1,280 degrees of freedom). Changes in hippocampal volume in untreated non-
depressed controls scanned at two different time points were also assessed to
estimate the variance associated with repeated measures over time.
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPT
Oltedal et al. 7
Methods and Materials
Participants. The clinical and demographic characteristics of the GEMRIC sample
are summarized in Table 1, as also detailed in (24). Data from 10 sites were
available, including n=281 patients (59.8% female, age 54.8 ± 16.4) and n=95
healthy controls (60% female, age 46.9 ± 14.6). Patients were scanned before and
after ECT, and controls were scanned at two time points without receiving ECT. Due
to some missing data points (e.g. follow up scan, number of ECTs, or depression
score), the sample sizes for the statistical models used to test for the effects of ECT
number or relationships with clinical outcome ranged from 250 to 268 patient
participants. ECT practice varied among contributing sites in terms of electrode
placement and/or stimulation parameters as detailed previously (24). Concurrent
psychotropic medications were used at most sites, as describe in Supplemental
information. To test for the effects of electrode placement, only patients that received
exclusively RUL (n=149) or BL (n=50; 10 bifrontal (BF), 40 bitemporal (BT))
treatment throughout all sessions of the ECT index series were included for analysis.
All sites contributing data received approval by their local ethical committees or
Institutional Review Board, and the centralized mega-analysis was approved by the
Regional Ethic Committee South-East in Norway (2013/1032 ECT and
Neuroradiology, June 1st 2015).
Image acquisition and post processing. The image processing methods have
been detailed previously (24). Briefly, T1-weighted MRI volumes with a minimal
resolution of 1.3 mm in any direction were acquired before and after (typically within
1-2 weeks) an ECT treatment series using 1.5T (1 site) or 3T (9 sites) scanners.
Raw structural MRI data from each site were uploaded to a common server and
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPT
Oltedal et al. 8
were analyzed together using the same preprocessing steps. During preprocessing,
images were corrected for scanner specific gradient non-linearity (25), registered to
a common atlas space and resampled to an isotropic 1 mm3 spatial resolution.
Further processing was performed by FreeSurfer version 5.3, and Quarc (26) was
used for unbiased estimation of hippocampal volume change. The automated
segmentation of FreeSurfer for hippocampal volume measurement has been shown
to be comparable to results from manual tracings (27-29). Depressive symptoms
were rated by the Montgomery-Åsberg Depression Rating Scale (MADRS). For sites
collecting only the 17- or 24-item Hamilton Depression Rating Scale (HAM-D), a
validated equation was used to convert HAM-D-17 to MADRS scores (30).
For all modes of electrode placement employed across sites, one of the electrodes
was placed over the right (non-dominant) hemisphere, hence the right hippocampus
was chosen for primary analysis to determine dose effects of repeated ECT
treatments and relationships with clinical response, weighting ECT session similarly
regardless of participant variations for electrode placement within or across sites.
The same effects were examined for the left hippocampus and results from these
analyses are provided in the Supplement. Follow-up analyses were performed to
examine the effects of BL and RUL electrode placement on both the right and left
hippocampus, excluding one patient that received left anterior right temporal (LART)
and patients who received a combination of RUL and BL during the index series.
Quality control of hippocampal segmentation was performed by procedures adapted
from the ENIGMA consortium (http://enigma.usc.edu/) (31).
Statistical analysis. Statistical analysis was performed with the R software
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPT
Oltedal et al. 9
package, version 3.3.1 (32). Slopes from linear models are reported with ± Standard
Error (SE) and all other results are reported as Mean ± Standard Deviation (SD).
Primary analyses addressed relationships between 1) the number of ECT sessions
and hippocampal volume change, and 2) hippocampal volume change and change
in MADRS score pre to post ECT using the General Linear Model (GLM). In a
subsample of patients receiving only BL or RUL ECT, effects of electrode placement
were additionally examined, and differences in slopes were tested using the function
linearHypothesis in R (car-package, version 2.1-6). To control for and evaluate non-
linear effects, the number of ECT sessions squared was included as a covariate. To
control for Age, Sex, Site, baseline hippocampal volume and baseline depression
score, these variables were included as covariates in the models as specified in the
Results. Considering our a priori hypotheses and the large amount of literature
showing changes in hippocampal volume with ECT (9), individual tests were
considered significant at a level of p < 0.01, corresponding to a Bonferroni correction
for 5 independent hypotheses. In the results figures, the regression lines (with 95%
confidence intervals shown as shaded areas) represent the relationships between
dependent and independent variables calculated without covariates. Cohen´s d for
volume change was calculated as mean change/SD. Finally, relationships between
volume change and number of ECT sessions were additionally examined in
responders (patients who showed >50% change in MADRS score over the course of
ECT, n = 150) versus non-responders (n = 98) using Welch Two Sample t-tests (two-
sided).
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPT
Oltedal et al. 10
Results
First, we tested whether volume change of the hippocampus is positively associated
with number of ECT sessions over time, including number of ECTs squared (to
estimate non-linear effects), Age, Sex, Site, baseline depression score, and baseline
hippocampal volume as covariates. For the right hippocampus, we found that the
linear component (slope) of volume change (%) versus number of ECTs was 0.28 ±
0.083, (t(225) = 3.35, p < 0.001). The square term was near significant -0.0048 ±
0.002, (t(225) = -1.94, p = 0.053), suggesting a sub-linear relationship (Figure 1A)
that reflects larger volume changes occur early in the ECT treatment series. When
comparing control subjects scanned at two distinct time points, no significant
changes in hippocampal volume were observed; mean 0.05 % ± 0.08, d = 0.06, n =
95; p = 0.54 (One Sample t-test). Results for the left hippocampus, which are
presented in the Supplement, showed similarly significant volume enlargement with
increasing number of ECT sessions. Mean volumes are provided in Table 1.
Next, we tested whether clinical outcome following ECT, measured using the
MADRS, is positively associated with change in right hippocampal volume, when
controlling for effects of Age, Sex, Site, baseline depression score and baseline
hippocampal volume. Contrary to our hypothesis that patients with greater clinical
response would exhibit larger volume increases, we found a negative relationship
(slope -1.0 ± 0.35, t(233) = -2.84, p < 0.005) (Figure 1B) indicating less change in
those with the greatest improvement. Separating patients based on the extent of
clinical response over the course of ECT, volume change (%) was 2.6 ± 2.0, d = 1.3
and 3.3 ± 1.7, d = 1.9 for responders (those with > 50% improvement in mood
scores) and non-responders, respectively (p = 0.009, Figure 1C). However, we also
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPT
Oltedal et al. 11
observed that the number of ECT sessions was associated with worse outcome
(Figure 1D and see Supplemental information) such that non-responders were
prescribed and received more sessions than responders (13.2 ± 4.7 versus 11.5 ±
5.3, t(232.11) = 2.74, p = 0.007). Thus, to control for differences in the length of
treatment for responsive versus non-responsive patients, the number of ECT
sessions was additionally included as covariate to the model addressing the
relationship between change in hippocampal volume and change in mood rating.
When additionally controlling for the number of ECT sessions, the slope of change in
MADRS score versus volume change remained negative, but was no longer
significant (-0.69 ± 0.38, t(225) = -1.83, p = 0.069). The effect size of hippocampal
volume change (partial eta squared) was 0.03 and 0.01 before and after adding
number of ECT sessions as a covariate. As shown in the Supplement, positive
relationships between left hippocampal volume enlargement and clinical change
were also absent. Follow-up analyses examining effects of ECT number and
relationships with clinical outcome in ECT responders and non-responders for both
the left and right hippocampus are also presented in the Supplemental information
(see Figure S1).
Finally, to investigate the effects electrode placement, we constructed separate
linear models for change in volume for the right and left hippocampus with separate
slopes for the number of RUL or BL ECT sessions, controlling for Age, Sex, Site,
baseline depression score and baseline hippocampal volume. For the right
hippocampus (Figure 2A), the slopes of volume change per ECT session for RUL
and BL electrode placement were both ~0.13, suggesting similar effects for number
of BL and RUL treatments. Change in volume (mean ± SD) was also similar for BL
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPT
Oltedal et al. 12
and RUL electrode placement, 3.0 ± 1.7%, d = 1.8 and 2.7 ± 2.0%, d = 1.4, p = 0.36,
t-test, respectively. For the left hippocampus (Figure 2B), the slope of volume
change (slope ± SE) versus number of treatments was steeper for BL (0.18 ± 0.03, p
= 1.9 x 10-7) than RUL (0.06 ± 0.04, p = 0.15) electrode placements (p = 0.007,
Linear hypothesis test). Change in left hippocampal volume was also greater for BL
with respect to RUL stimulation (BL: 3.3 ± 2.2%, d = 1.5; RUL: 1.6 ± 2.1%, d = 0.8, p
= 1.5 x 10-5, t-test). The effect of electrode placement on the left hippocampal
volume change was further confirmed by a number of ECTs-by-electrode placement
interaction (p = 0.007) in a model of left hippocampal volume change versus number
of ECTs where electrode placement was included as a separate covariate (see
Supplemental Information, Model 2c).
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPT
Oltedal et al. 13
Discussion
Including the largest sample of patients with ECT studied with neuroimaging
methods to date, our findings showed a highly significant number of ECT session
dose-dependent biological effect of ECT on hippocampal volume. We also showed
that electrode placement differentially affects the extent of volume change in the right
and left hippocampus. Specifically, BL stimulation accounts for similar changes in
volume for both the right and left hippocampus, but RUL stimulation lead to more
focal effects in the right hippocampus. However, contrary to our expectations, we
also found that volume enlargement of the hippocampus is not significantly related to
treatment outcome. Instead, results showed a negative relationship between
hippocampal volume and symptom improvement such that individuals with greater
hippocampal enlargement tend to have less response. However, patients with poor
response received more treatments, and this negative relationship was not
significant when the number of ECT sessions were taken into account. This finding
represents a major deviation from the common assumption in the field of a positive
association between ECT-induced volume enlargement and clinical improvement.
Rather, results indicate that gross volume increase of the hippocampus by itself is
not a meaningful biomarker for positive therapeutic response.
Findings from this study showed that ECT dose parameters including the number of
ECT sessions received and the location of electrode placement modulated the
magnitude and hemispheric specificity of hippocampal volume change. Here, results
demonstrated a clear and dose-dependent effect of number of ECT sessions on
hippocampal volume in both the right and left hemispheres. Further, RUL and BL
ECT showed differential effects on volume change in the left and right hippocampus.
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPT
Oltedal et al. 14
Existing data supports that the antidepressant efficacy and cognitive side effects of
ECT are influenced by electrode position as well as other stimulus parameters (17,
33, 34). Designed to reduce cognitive side effects, with RUL electrode placement,
electrical stimulation is focused away from the dominant (left) hemisphere (35). In
contrast, the right side of the brain is targeted by both RUL and BL electrode
placements. Hence, if the electrical stimulus is modulating the volume change, a
clear difference in volumetric effect of RUL versus BL stimulation for the left
hippocampus is expected. In line with this hypothesis, and computational modelling
results showing more prominent electric field increases in the right hemisphere for
RUL ECT and in both hemispheres for BL ECT (20, 21), our results show volume
increases are greater in the right hippocampus for RUL, while BL ECT leads to
similar volume increases in both hemispheres (Figure 2).
Though we have shown that hippocampal volume enlargement is influenced by ECT
dose parameters, the clinical relevance of these changes remains unclear. ECT-
induced volume enlargement of the hippocampus (4-8) has led to the suggestion that
treatment-related neuroplasticity may underlie symptom improvement (12). From a
mechanistic perspective, stress in combination with genetic or epigenetic factors
may reduce neurogenesis and precipitate a depressive episode, and antidepressant
therapies (such as ECT) might work through restoration of the basal rate of
neurogenesis in the hippocampal dentate gyrus (11). Since both left (Figure S1B and
D) and right (Figure 1) hippocampal volume change relates to the number of ECT
treatments received, but does not positively associate with clinical outcome,
enlargement of the hippocampus may be an epiphenomenon of ECT. Overall
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPT
Oltedal et al. 15
enlargement of hippocampal volume observed with ECT may thus relate to seizure
therapy itself rather than to the therapeutic effects of treatment.
Our results have important implications for treatment management and raise several
questions and challenges relevant to understanding the neurobiological
underpinnings of ECT. It is a common experience among ECT-practitioners that the
patients with the highest depression scores tend to be the ones with the higher
response rates (36), and often these patients respond quickly. At the same time,
longer depressive episodes and medication failure at baseline are indicators of poor
response to ECT (37). The number of treatments prescribed is typically based on
clinically determined response, and patients with modest response are thus more
likely to receive a larger number of ECT sessions in the index series (16). However,
while the biological effects of ECT may be expected to relate to the number of
treatments received, as shown for growth of the hippocampus, there is not an
apparent parallel regarding improvement in depression score (Figure 1D).
It is conceivable that several different biological processes impact ECT clinical
response and these might or might not overlap with the biological manifestations of
seizure therapy itself. Animal studies support that in addition to neurogenesis,
multiple other neurophysiological and neuroplastic changes occur following
electroconvulsive shock (ECS). Thus, it is possible that particular micro-
environmental events may influence the overall macroscopic structure of the
hippocampus, while separate or concurrent processes constitute the mechanisms
underlying antidepressant response. For example, changes in cellular or synaptic
density and intra/extracellular fluid might impact gross changes in hippocampal
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPT
Oltedal et al. 16
volume. Animal models have shown dose-dependent increase in markers of
hippocampal neural, glial and endothelial cell proliferation and density following ECS
(15, 38-40) that may result in an absolute increase in the number of synapses or
specific cell types (41). Notably, a dissociation between neural changes and
behavior was reported in a recent animal model study, where ECT was shown to
stimulate neurogenesis, but the number of new neurons did not predict the extent of
behavioral outcome (42). These results are compatible with our findings with respect
to the absence of clinical response relationships. At the same time, hippocampal
volume may be influenced by fluid content, which may vary as a consequence of
increased vascularization (43) and blood flow (44, 45), or inflammation (46-48) as
supported by an observed ECS upregulation of markers for microglia (49, 50).
Other molecular effects, not necessarily independent, may relate more directly to
antidepressant response. For example, ECS is also shown to modulate
monoaminergic neurotransmission (51), as similar to standard antidepressant
treatment. Increased expression of brain-derived neurotrophic factor (BDNF) (52, 53)
and vascular endothelial growth factor (VEGF) (54) are also reported with ECS or
ECT in humans, and have been linked to changes in behavior (52, 55). Further, ECS
elicits a number of hippocampal epigenetic modifications, including GADD45B-
dependent DNA demethylation (56), and the alteration of histone and DNA modifying
enzymes (57), which may influence structural neuroplasticity at both the macro and
micro-scale.
It is also possible that neurogenesis or other neurotrophic or neurophysiological
events induced by ECT may precede or lag behind clinical response. Further,
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPT
Oltedal et al. 17
variations in the morphology of different regions of the hippocampus (for example,
the dentate gyrus or the anterior hippocampus with more connections to neural
circuits associated with mood regulation and emotional behavior) may be more
sensitive to ECT outcome. For example, analyses of change in hippocampal shape
with ECT have indicated greater regional changes in the right anterior hippocampus
(12), as well as changes specific to particular hippocampal subfields (7). A recent
study in 24 subjects also suggested that volumes of hippocampal subfields at
baseline could predict response to ECT treatment (58), however this finding needs
replication in larger samples.
Our study has some limitations, most notably the design is retrospective (e.g. no a
priori standardization of MR protocols or depression scoring) and assessments were
limited to before- and after treatment. Further, the design was naturalistic, so
patients who remained unresponsive were prescribed a greater number of ECT
sessions on average. Other unknown moderators or speed of response, which can
impact clinical decisions regarding the number of treatments prescribed (59), remain
similarly unaccounted for. For example, other stimulation parameters such as pulse
width and frequency and seizure threshold may also impact neural changes.
However, since these parameters varied across sites, including during the ECT
treatment series for individual patients, they were not investigated. Animal studies
have also shown that both ECS, and to a lesser extent, chronic antidepressant
treatment impact neurogenesis in the rat hippocampus (38). It is thus possible that
the continuation of psychotropic medication during ECT might impact hippocampal
structure. However, follow-up analysis revealed the extent of volume change was
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPT
Oltedal et al. 18
similar for participants tapered off all antidepressants, benzodiazepines and
anticonvulsants during ECT (Figure S2).
Cognitive side effects remain a fundamental concern in ECT practice, and were not
examined in this study and thus warrant future research. Future studies would also
benefit from including repeated assessments at multiple time points throughout
treatment to allow examination of the trajectories and speed of change, and explore
ways of subgrouping depressed individuals, possibly by identifying biological
subtypes (60). Implementing machine-learning approaches, with a goal of identifying
individuals that are likely to respond to ECT (61), and investigations using higher
resolution imaging approaches to investigate sub-regions of the hippocampus (58)
may also advance the field. Another avenue of future research would be studies with
standardized ECT protocols across all participants to reduce confounds and increase
the power of the designs to identify moderators conclusively. New approaches are
needed to identify biomarkers that can explain and predict the clinical effect of ECT,
separate from seizure or other procedural effects, which also may inform other
antidepressant treatments.
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPT
Oltedal et al. 19
Acknowledgements
This study is supported by Western Norway Regional Health Authority, Haukeland
University Hospital and the University of Bergen, Norway. LO acknowledge support
from the Fulbright Program. Individual sites acknowledge funding from: The
Lundbeck Foundation (Denmark). The German Research Foundation (DFG, grant
FOR2107 DA1151/5-1 to UD; SFB-TRR58, Project C09 to UD) and the
Interdisciplinary Center for Clinical Research (IZKF) of the medical faculty of Münster
(grant Dan3/012/17 to UD); DFG, grant FOR2107 HA7070/2-2 to TH; Innovative
Medizinische Forschung (IMF, RE111722 and RE111604 to RR), and the National
Institute of Mental Health (MH092301, MH110008 and MH102743).
Some preliminary findings were presented to the Society of Biological Psychiatry
meeting in 2017, abstract 447 (Oltedal, L., et al. (2017). Biological Psychiatry 81(10,
Supplement): S182-S183.)
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPT
Oltedal et al. 20
Financial Disclosures
Anders M. Dale is a Founder of and holds equity in CorTechs Labs, Inc, and serves
on its Scientific Advisory Board. He is also a member of the Scientific Advisory Board
of Human Longevity, Inc. (HLI), and receives funding through research agreements
with General Electric Healthcare (GEHC). The terms of these arrangements have
been reviewed and approved by the University of California, San Diego in
accordance with its conflict of interest policies.
All other authors report no biomedical financial interests or potential conflicts of
interest.
Author contributions
LO wrote the first draft and coordinated the work. LO, UK, HB, KJE, OBP, MBJ, CCA
and AMD contributed in planning and/or design the project. LO, UK, KJO, VJE, MBJ,
LGH, KLN, CCA, AD, MLS, MLO, LE, MV, PS, PvE, IT, MA, RR, TH, UD, AA, RE
contributed data. LO, HB, AMD, KLN, AD, MLS, MLO, LE, MV, PS, MA, RR, AA, GH,
RE and CCA contributed in processing and/or analysis/interpretation of data. All
authors contributed to manuscript revisions and approved of the final version.
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPT
Oltedal et al. 21
References
1. Friedrich MJ (2017): Depression Is the Leading Cause of Disability Around the World.
JAMA. 317:1517.
2. Insel TR, Wang PS (2009): The STAR*D trial: revealing the need for better treatments.
Psychiatr Serv. 60:1466-1467.
3. Carney S, Cowen P, Dearness K, Eastaugh J (2003): Efficacy and safety of
electroconvulsive therapy in depressive disorders: a systematic review and meta-analysis.
Lancet. 361:799-808.
4. Nordanskog P, Dahlstrand U, Larsson MR, Larsson EM, Knutsson L, Johanson A
(2010): Increase in hippocampal volume after electroconvulsive therapy in patients with
depression: a volumetric magnetic resonance imaging study. J Ect. 26:62-67.
5. Tendolkar I, van Beek M, van Oostrom I, Mulder M, Janzing J, Voshaar RO, et al.
(2013): Electroconvulsive therapy increases hippocampal and amygdala volume in therapy
refractory depression: a longitudinal pilot study. Psychiatry research. 214:197-203.
6. Dukart J, Regen F, Kherif F, Colla M, Bajbouj M, Heuser I, et al. (2014):
Electroconvulsive therapy-induced brain plasticity determines therapeutic outcome in mood
disorders. Proc Natl Acad Sci U S A. 111:1156-1161.
7. Abbott CC, Jones T, Lemke NT, Gallegos P, McClintock SM, Mayer AR, et al. (2014):
Hippocampal structural and functional changes associated with electroconvulsive therapy
response. Translational psychiatry. 4:e483.
8. Ota M, Noda T, Sato N, Okazaki M, Ishikawa M, Hattori K, et al. (2015): Effect of
electroconvulsive therapy on gray matter volume in major depressive disorder. J Affect
Disord. 186:186-191.
9. Wilkinson S, T., Sanacora G, Bloch MH (2017): Hippocampal Volume Changes
Following Electroconvulsive Therapy: A Systematic Review and Meta-analysis. Biol
Psychiatry. 2:327-335.
10. Miller BR, Hen R (2015): The current state of the neurogenic theory of depression
and anxiety. Curr Opin Neurobiol. 30:51-58.
11. Jacobs BL, van Praag H, Gage FH (2000): Adult brain neurogenesis and psychiatry: a
novel theory of depression. Mol Psychiatry. 5:262-269.
12. Joshi SH, Espinoza RT, Pirnia T, Shi J, Wang Y, Ayers B, et al. (2016): Structural
Plasticity of the Hippocampus and Amygdala Induced by Electroconvulsive Therapy in Major
Depression. Biol Psychiatry. 79:282-292.
13. Eriksson PS, Perfilieva E, Bjork-Eriksson T, Alborn AM, Nordborg C, Peterson DA, et al.
(1998): Neurogenesis in the adult human hippocampus. Nat Med. 4:1313-1317.
14. Schmaal L, Veltman DJ, van Erp TG, Samann PG, Frodl T, Jahanshad N, et al. (2016):
Subcortical brain alterations in major depressive disorder: findings from the ENIGMA Major
Depressive Disorder working group. Mol Psychiatry. 21:806-812.
15. Madsen TM, Treschow A, Bengzon J, Bolwig TG, Lindvall O, Tingstrom A (2000):
Increased neurogenesis in a model of electroconvulsive therapy. Biol Psychiatry. 47:1043-
1049.
16. Brus O, Cao Y, Gustafsson E, Hulten M, Landen M, Lundberg J, et al. (2017): Self-
assessed remission rates after electroconvulsive therapy of depressive disorders. Eur
Psychiatry. 45:154-160.
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPT
Oltedal et al. 22
17. Kolshus E, Jelovac A, McLoughlin DM (2017): Bitemporal v. high-dose right unilateral
electroconvulsive therapy for depression: a systematic review and meta-analysis of
randomized controlled trials. Psychol Med. 47:518-530.
18. Radman T, Lisanby SH (2017): New directions in the rational design of electrical and
magnetic seizure therapies: individualized Low Amplitude Seizure Therapy (iLAST) and
Magnetic Seizure Therapy (MST). Int Rev Psychiatry. 29:63-78.
19. Tor PC, Bautovich A, Wang MJ, Martin D, Harvey SB, Loo C (2015): A Systematic
Review and Meta-Analysis of Brief Versus Ultrabrief Right Unilateral Electroconvulsive
Therapy for Depression. J Clin Psychiatry. 76:e1092-1098.
20. Bai S, Galvez V, Dokos S, Martin D, Bikson M, Loo C (2017): Computational models of
Bitemporal, Bifrontal and Right Unilateral ECT predict differential stimulation of brain
regions associated with efficacy and cognitive side effects. Eur Psychiatry. 41:21-29.
21. Deng ZD, Lisanby SH, Peterchev AV (2011): Electric field strength and focality in
electroconvulsive therapy and magnetic seizure therapy: a finite element simulation study. J
Neural Eng. 8:016007.
22. Jorgensen A, Magnusson P, Hanson LG, Kirkegaard T, Benveniste H, Lee H, et al.
(2016): Regional brain volumes, diffusivity, and metabolite changes after electroconvulsive
therapy for severe depression. Acta Psychiatr Scand. 133:154-164.
23. Bouckaert F, Dols A, Emsell L, De Winter FL, Vansteelandt K, Claes L, et al. (2016):
Relationship Between Hippocampal Volume, Serum BDNF, and Depression Severity
Following Electroconvulsive Therapy in Late-Life Depression. Neuropsychopharmacology.
41:2741-2748.
24. Oltedal L, Bartsch H, Sorhaug OJ, Kessler U, Abbott C, Dols A, et al. (2017): The Global
ECT-MRI Research Collaboration (GEMRIC): Establishing a multi-site investigation of the
neural mechanisms underlying response to electroconvulsive therapy. Neuroimage Clin.
14:422-432.
25. Jovicich J, Czanner S, Greve D, Haley E, van der Kouwe A, Gollub R, et al. (2006):
Reliability in multi-site structural MRI studies: effects of gradient non-linearity correction on
phantom and human data. NeuroImage. 30:436-443.
26. Holland D, Dale AM (2011): Nonlinear registration of longitudinal images and
measurement of change in regions of interest. Med Image Anal. 15:489-497.
27. Fischl B, Salat DH, Busa E, Albert M, Dieterich M, Haselgrove C, et al. (2002): Whole
brain segmentation: automated labeling of neuroanatomical structures in the human brain.
Neuron. 33:341-355.
28. Morey RA, Petty CM, Xu Y, Hayes JP, Wagner HR, 2nd, Lewis DV, et al. (2009): A
comparison of automated segmentation and manual tracing for quantifying hippocampal
and amygdala volumes. NeuroImage. 45:855-866.
29. Sanchez-Benavides G, Gomez-Anson B, Sainz A, Vives Y, Delfino M, Pena-Casanova J
(2010): Manual validation of FreeSurfer's automated hippocampal segmentation in normal
aging, mild cognitive impairment, and Alzheimer Disease subjects. Psychiatry research.
181:219-225.
30. Heo M, Murphy CF, Meyers BS (2007): Relationship between the Hamilton
Depression Rating Scale and the Montgomery-Asberg Depression Rating Scale in depressed
elderly: a meta-analysis. Am J Geriatr Psychiatry. 15:899-905.
31. Stein JL, Medland SE, Vasquez AA, Hibar DP, Senstad RE, Winkler AM, et al. (2012):
Identification of common variants associated with human hippocampal and intracranial
volumes. Nat Genet. 44:552-561.
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPT
Oltedal et al. 23
32. R Core Team (2016): R: A language and environment for statistical computing.
Vienna, Austria. URL https://www.R-project.org/: R Foundation for Statistical Computing.
33. Spellman T, Peterchev AV, Lisanby SH (2009): Focal electrically administered seizure
therapy: a novel form of ECT illustrates the roles of current directionality, polarity, and
electrode configuration in seizure induction. Neuropsychopharmacology. 34:2002-2010.
34. Sackeim HA, Prudic J, Devanand DP, Nobler MS, Lisanby SH, Peyser S, et al. (2000): A
prospective, randomized, double-blind comparison of bilateral and right unilateral
electroconvulsive therapy at different stimulus intensities. Arch Gen Psychiatry. 57:425-434.
35. d'Elia G (1970): Unilateral electroconvulsive therapy. Acta Psychiatr Scand Suppl.
215:1-98.
36. Nordenskjold A, von Knorring L, Engstrom I (2012): Predictors of the short-term
responder rate of Electroconvulsive therapy in depressive disorders--a population based
study. BMC Psychiatry. 12:115.
37. Haq AU, Sitzmann AF, Goldman ML, Maixner DF, Mickey BJ (2015): Response of
depression to electroconvulsive therapy: a meta-analysis of clinical predictors. J Clin
Psychiatry. 76:1374-1384.
38. Malberg JE, Eisch AJ, Nestler EJ, Duman RS (2000): Chronic antidepressant treatment
increases neurogenesis in adult rat hippocampus. J Neurosci. 20:9104-9110.
39. Perera TD, Coplan JD, Lisanby SH, Lipira CM, Arif M, Carpio C, et al. (2007):
Antidepressant-induced neurogenesis in the hippocampus of adult nonhuman primates. J
Neurosci. 27:4894-4901.
40. Scott BW, Wojtowicz JM, Burnham WM (2000): Neurogenesis in the dentate gyrus of
the rat following electroconvulsive shock seizures. Exp Neurol. 165:231-236.
41. Chen F, Madsen TM, Wegener G, Nyengaard JR (2009): Repeated electroconvulsive
seizures increase the total number of synapses in adult male rat hippocampus. Eur
Neuropsychopharmacol. 19:329-338.
42. Olesen MV, Wortwein G, Folke J, Pakkenberg B (2017): Electroconvulsive stimulation
results in long-term survival of newly generated hippocampal neurons in rats. Hippocampus.
27:52-60.
43. Hellsten J, Wennstrom M, Bengzon J, Mohapel P, Tingstrom A (2004):
Electroconvulsive seizures induce endothelial cell proliferation in adult rat hippocampus.
Biol Psychiatry. 55:420-427.
44. Milo TJ, Kaufman GE, Barnes WE, Konopka LM, Crayton JW, Ringelstein JG, et al.
(2001): Changes in regional cerebral blood flow after electroconvulsive therapy for
depression. J Ect. 17:15-21.
45. Takano H, Motohashi N, Uema T, Ogawa K, Ohnishi T, Nishikawa M, et al. (2011):
Differences in cerebral blood flow between missed and generalized seizures with
electroconvulsive therapy: a positron emission tomographic study. Epilepsy Res. 97:225-228.
46. Yrondi A, Sporer M, Peran P, Schmitt L, Arbus C, Sauvaget A (2018): Electroconvulsive
therapy, depression, the immune system and inflammation: A systematic review. Brain
Stimul. 11:29-51.
47. van Buel EM, Patas K, Peters M, Bosker FJ, Eisel UL, Klein HC (2015): Immune and
neurotrophin stimulation by electroconvulsive therapy: is some inflammation needed after
all? Translational psychiatry. 5:e609.
48. Fluitman SB, Heijnen CJ, Denys DA, Nolen WA, Balk FJ, Westenberg HG (2011):
Electroconvulsive therapy has acute immunological and neuroendocrine effects in patients
with major depressive disorder. J Affect Disord. 131:388-392.
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPT
Oltedal et al. 24
49. Jansson L, Wennstrom M, Johanson A, Tingstrom A (2009): Glial cell activation in
response to electroconvulsive seizures. Prog Neuropsychopharmacol Biol Psychiatry.
33:1119-1128.
50. Wennstrom M, Hellsten J, Tingstrom A (2004): Electroconvulsive seizures induce
proliferation of NG2-expressing glial cells in adult rat amygdala. Biol Psychiatry. 55:464-471.
51. Baldinger P, Lotan A, Frey R, Kasper S, Lerer B, Lanzenberger R (2014):
Neurotransmitters and electroconvulsive therapy. J Ect. 30:116-121.
52. Polyakova M, Schroeter ML, Elzinga BM, Holiga S, Schoenknecht P, de Kloet ER, et al.
(2015): Brain-Derived Neurotrophic Factor and Antidepressive Effect of Electroconvulsive
Therapy: Systematic Review and Meta-Analyses of the Preclinical and Clinical Literature.
PLoS One. 10:e0141564.
53. Nibuya M, Morinobu S, Duman RS (1995): Regulation of BDNF and trkB mRNA in rat
brain by chronic electroconvulsive seizure and antidepressant drug treatments. J Neurosci.
15:7539-7547.
54. Warner-Schmidt JL, Madsen TM, Duman RS (2008): Electroconvulsive seizure
restores neurogenesis and hippocampus-dependent fear memory after disruption by
irradiation. Eur J Neurosci. 27:1485-1493.
55. Minelli A, Zanardini R, Abate M, Bortolomasi M, Gennarelli M, Bocchio-Chiavetto L
(2011): Vascular Endothelial Growth Factor (VEGF) serum concentration during
electroconvulsive therapy (ECT) in treatment resistant depressed patients. Prog
Neuropsychopharmacol Biol Psychiatry. 35:1322-1325.
56. Jun H, Mohammed Qasim Hussaini S, Cho CH, Welby J, Jang MH (2015): Gadd45b
Mediates Electroconvulsive Shock Induced Proliferation of Hippocampal Neural Stem Cells.
Brain Stimul. 8:1021-1024.
57. Pusalkar M, Ghosh S, Jaggar M, Husain BF, Galande S, Vaidya VA (2016): Acute and
Chronic Electroconvulsive Seizures (ECS) Differentially Regulate the Expression of Epigenetic
Machinery in the Adult Rat Hippocampus. Int J Neuropsychopharmacol. 19.
58. Cao B, Luo Q, Fu Y, Du L, Qiu T, Yang X, et al. (2018): Predicting individual responses
to the electroconvulsive therapy with hippocampal subfield volumes in major depression
disorder. Sci Rep. 8:5434.
59. Cinar S, Oude Voshaar RC, Janzing JG, Birkenhager TK, Buitelaar JK, van den Broek
WW (2010): The course of depressive symptoms in unipolar depressive disorder during
electroconvulsive therapy: a latent class analysis. J Affect Disord. 124:141-147.
60. Drysdale AT, Grosenick L, Downar J, Dunlop K, Mansouri F, Meng Y, et al. (2017):
Resting-state connectivity biomarkers define neurophysiological subtypes of depression.
Nat Med. 23:28-38.
61. Redlich R, Opel N, Grotegerd D, Dohm K, Zaremba D, Burger C, et al. (2016):
Prediction of Individual Response to Electroconvulsive Therapy via Machine Learning on
Structural Magnetic Resonance Imaging Data. JAMA Psychiatry. 73:557-564.
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPT
Oltedal et al. 25
Figure Legends
Figure 1 | Differential effect of ECT on hippocampal volume and clinical
outcome. A, Scatter plot of volume change of the right hippocampus, computed as
(posttreatment – pretreatment score)/pretreatment score x 100, versus number of
ECTs; n = 241. Slope (controlling for number of ECTs squared, Age, Sex, Site,
baseline depression score, and baseline hippocampal volume), 0.28 ± 0.08, (t(225) =
3.35, p < 0.001. B, Scatter plot of change in MADRS score, computed as
pretreatment – posttreatment score, versus volume change of the right
hippocampus; n = 248. Slope (controlling for Age, Sex, Site, baseline depression
score and baseline hippocampal volume), -1.0 ± 0.35, t(233) = -2.84, p < 0.005. C,
Boxplot comparing volume change of the right hippocampus in non-responders
(MADRS reduction < 50%) versus responders (MADRS reduction > 50%), n = 248,
t(234.13) = 2.62, p = 0.009. D, Scatter plot of change in MADRS score versus
number of ECTs; n = 268. Slope (controlling for age, Sex and Site), - 0.28 ± 0.16,
t(256) = -1.80, p = 0.074). Non-responders received more ECT sessions (13.2 ± 4.7
versus 11.5 ± 5.3, t(232.11) = 2.74, p = 0.007) than responders.
Figure 2 | Effect of electrode placement on change in left and right
hippocampal volume. A. Changes in right hippocampal volume per number of ECT
sessions for bilateral (BL, dashed line) and right unilateral (RUL, solid line) electrode
placement. Both slope and change in volume was similar for BL and RUL ECT
(slope: both ~.13; BL volume increase: 3.0 ± 1.7%, RUL volume increase: 2.7 ±
2.0%). B. Changes in left hippocampal volume per number of ECT sessions for BL
(dashed line) and RUL (solid line) electrode placement. Slope was steeper and
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPT
Oltedal et al. 26
volume change was greater for BL (slope: 0.18 ± 0.03; volume increase: 3.3 ± 2.2%)
than RUL (slope: 0.06 ± 0.04; volume increase: 1.6 ± 2.1%) stimulation.
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPTCharacteristics of subjects Mean SD n#
Controls Age 46.9 14.6 95 Baseline right hippocampal volume (mm3) 4052.5 446.2 95 Change in right hippocampal volume (%) 0.05 0.8 95 Baseline left hippocampal volume (mm3) 3948.0 444.3 95 Change in left hippocampal volume (%) 0.01 0.7 95 Baseline intracranial volume (cm3) 1520.2 179.2 95 Patients Age 54.9 16.4 281 Baseline right hippocampal volume (mm3) 3774.1 588.3 254$
Change in right hippocampal volume (%) 2.9 1.9 250^ Baseline left hippocampal volume (mm3) 3657.9 561.0 254$ Change in left hippocampal volume (%) 2.2 2.3 250^ Baseline intracranial volume (cm3) 1505.9 175.6 254$ Baseline depression score 33.3 8.2 279 Post treatment depression score 15.0 11.0 277 Duration of episode (months) 20.1 31.6 158 Number of ECTs, total 12.0 5.2 273* Number of ECTs, BL only 14.6 7.5 50 Number of ECTs, RUL only 10.9 3.6 149 Number of ECTs, responders 11.5 5.3 166 Number of ECTs, non-responders 13.2 4.7 102
Table 1 The number of subjects (#) vary because of missing data for some variables. Information about number of ECTs (*) was missing for 8 subjects; some subjects received more than one form of lead placement and one subject also received LART stimulation. A total of 27 subjects ($) were missing MRI at either before or after treatment (baseline volume is not reported for these) and 4 subjects failed automated processing of volume change (^).
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPT
−5
0
5
10
0 10 20 30 40Number of ECTs
Volu
me
chan
ge, r
ight
hip
poca
mpu
s (%
)Treatment & volume change
0
20
40
0 4 8Volume change, right hippocampus (%)
MAD
RS
chan
ge (p
re −
pos
t)
Volume change & outcome
0
5
NR RNon−responders (NR) and responders (R)
Volu
me
chan
ge, r
ight
hip
poca
mpu
s (%
)
Outcome & volume change
0
20
40
0 10 20 30 40Number of ECTs
MAD
RS
chan
ge (p
re −
pos
t)Treatment & outcome
A B
C D
Figure 1
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPT
−5
0
5
10
0 10 20 30 40Number of ECTs
Volu
me
chan
ge (%
)
Stimulation typeBL
RUL
Right Hippocampus
−5
0
5
10
0 10 20 30 40Number of ECTs
Volu
me
chan
ge (%
)
Stimulation typeBL
RUL
Left Hippocampus
Figure 2
A B
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPTOltedal et al. Supplement
1
Volume of the Human Hippocampus and Clinical Response Following Electroconvulsive Therapy
Supplemental Information
Volume change and clinical outcome for the left hippocampus
Volume growth was associated with the number of ECT sessions for both the right (see main
text) and the left hippocampus (t(225) = 3.93, p < 0.001). Controlling for the number of ECT
sessions, relationships between change in clinical response and change in hippocampal
volume were not significant for the right (see main text) or left hippocampus (t(225) = -1.50,
p = 0.14).
Volume change for ECT responders and non-responders
Although the number of ECT sessions received for individual patients was confounded by
the extent of their clinical response, understanding how clinical response relates to length of
treatment provides further insight. Specifically, a model of change in MADRS score as a
function of number of treatments showed a negative association F(1,266) = 7.96, slope =
-0.41, p = 0.005, R2 = 0.03 (Figure S1D), although after correcting for Age, Sex and Site this
was not significant (slope = -0.28, t(256) = -1.80, p = 0.074). Hence, follow up analyses were
performed to further clarify the effects of volume change and ECT session number in ECT
responders and non-responders. First, we tested for the effect of response group
(responder, non-responder) on hippocampal volume change, correcting for number of ECTs
and number of ECTs squared, and found that the effect of group was not significant for the
right (-0.35, t(236) = -1.41, p = 0.16), or the left (-0.50, t(236) = -1.78, p = 0.08) hippocampus
(Figure S1A and B, respectively). Also, looking at responders separately from non-
responders (Figure S1C and D) did not reveal positive relationships between volume change
and clinical outcome (change in MADRS score) for the right or left hippocampus. Similarly,
there was no positive association between number of ECTs and outcome for responders
separated from non-responders (Figure S1E). Finally, to completely exclude variance
associated with number of treatments, we used data from a subset of patients that received
exactly 12 ECT sessions (the mode number of treatments). Again, a model of change in
MADRS score as a function of volume change did not yield a significant relationship (F(1,45)
= 0.006, p = 0.94, R2 = 0.0001) (Figure S1F).
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPTOltedal et al. Supplement
2
Figure S1. Effect of ECT on hippocampal volume and clinical outcome for responders and non-responders. A, Scatter plot of volume change for the right hippocampus versus number of ECTs in responders (blue) and non-responders (red); n = 241. B, Scatter plot of volume change for the left hippocampus versus number of ECTs in responders (blue) and non-responders (red); n = 241. C, Scatter plot of change in MADRS score versus volume change for the right hippocampus in responders (blue) and non-responders (red); n = 247. D, Scatter plot of change in MADRS score versus volume change for the left hippocampus in responders (blue) and non-responders (red); n = 247. E, Scatter plot of change in MADRS score versus number of ECTs in responders (blue) and non-responders (red); n = 267. F, Scatter plot of change in MADRS versus number of ECTs including patients only receiving ≤ 12 sessions classified as responders (blue) and non-responders (red); n = 171.
0
20
40
0 10 20 30 40Number of ECTs
MA
DR
S c
han
ge (
pre
− p
ost)
response
NR
R
Treatment & outcome
0
20
40
0 10 20 30 40Number of ECTs
MA
DR
S c
han
ge (
pre
− p
ost)
response
NR
R
Treatment & outcome
0
20
40
0 4 8Volume change, right hippocampus (%)
MA
DR
S c
hang
e (p
re −
pos
t)
response
NR
R
Volume change & outcome
0
20
40
0 4 8Volume change, left hippocampus (%)
MA
DR
S c
hang
e (p
re −
pos
t)response
NR
R
Volume change & outcome
−5
0
5
10
0 10 20 30 40Number of ECTs
Vol
ume
cha
nge,
left
hip
poca
mpu
s (%
)
response
NR
R
Treatment & volume change
−5
0
5
10
0 10 20 30 40Number of ECTs
Vol
ume
cha
nge,
rig
ht h
ipp
oca
mpu
s (%
)
response
NR
R
Treatment & volume changeA B
C D
E F
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPTOltedal et al. Supplement
3
Supplementary Methods
Medications: Concurrent psychotropic medications were used at most sites, but there was
virtually no change in medication during treatment (number of patients receiving each
medication at baseline: antidepressants (n = 115), antipsychotics (n = 93), lithium (n = 0),
other mood stabilizer (n = 35) or benzodiazepines (n = 56)). Information about dose was not
available. Volume change of the right hippocampus did not differ between one site that
tapered patients off all antidepressants, benzodiazepines and anticonvulsants (Figure S2)
compared to all other sites (p = 0.27).
Electrode placement and clinical outcome: No significant difference for electrode
placement was found when modelling separate slopes for RUL and BL stimulation (Figure
S3).
0
20
40
0 10 20 30 40
Number of ECTs
MA
DR
S c
hang
e (p
re−
post
)
Stimulation type
BL
RUL
Lead placement and outcome
−5
0
5
10
0 10 20 30 40Number of ECTs
Volu
me
chan
ge, r
ight
hip
poc
ampus
(%
)
Treatment & volume change
Figure S2. Scatter plot of volume change of the
right hippocampus versus number of ECTs for
the site that tapered patients off all medications
before treatment. The mean volume change at this
site was 2.6%, which did not differ from all other sites
(2.9%), p = 0.27 t-test.
Figure S3. Treatment and clinical outcome for RUL
and BL. Scatter plot of change in MADRS score
versus number of ECTs for BL and RUL electrode
placements. In a model with BL and RUL as separate
predictors (controlling for age, sex, site, and baseline
depression score), the slopes were -0.27 and -0.53 for
BL and RUL stimulation, respectively. These slopes
were not significantly different (p = 0.25, Linear
Hypothesis test).
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPTOltedal et al. Supplement
4
Supplementary results, complete statistical models
The full linear models for our main analysis are provided below. The model name
corresponds to results presented in the Figures and in the main text.
Model 1a
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPTOltedal et al. Supplement
5
Model 1b
Model 1b including the number of ECT sessions
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPTOltedal et al. Supplement
6
Model 1d
Model 2a
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPTOltedal et al. Supplement
7
Model 2b
Model 2c